Saturday, December 14, 2024

Should You use a Dynamic Process with Guardrails to Keep Spending on Track in Retirement? Yes, But Not with a Monte Carlo Model Typically Used by Financial Advisors

This post is a follow-up to our post of March 7, 2021 and several other of our posts and articles noting that a model like the Actuarial Financial Planner (AFP) is a better model to use when employing a dynamic process to keep spending on track during retirement. 

In his December 11, 2024 Kitces.com article, Justin Fitzpatrick of Income Lab, says

“CMAs [Capital Market Assumptions], even perfectly accurate ones, can't predict exactly how much someone will be able to spend in retirement. But clients still want answers to the basic question, "How much can I afford to spend?" The key to addressing this dilemma is to recognize that retirement income planning doesn't require a one-time-only 'set it and forget it' decision. Instead, the ongoing process of guiding someone through retirement involves revisiting this question as the years – and even decades – progress.

When we structure retirement planning as ongoing guidance, assuming that clients are willing and able to make adjustments, the predictive weakness of CMAs becomes less problematic. Structuring retirement income planning in this way involves the following steps [Process]:

  1. Build a full picture of client resources, needs, and preferences.
  2. Using a reasonable model of the world, including CMAs, estimate a range of spending levels that may be possible for the client(s) today.
  3. Find an optimal spending level within this range that aligns with client risk preferences, balancing their ability and willingness to take on the risk of future cuts in income with their desire to spend and live life well. This becomes the answer to the question, "How much can I spend?" Advisors who are especially concerned that their return assumptions may be too high can target a spending level with a lower risk of overspending.
  4. Set 'guardrails' to signal when clients can spend more (because their risk of underspendingis too high) or less (because their risk of overspending is too high). These guardrails will address the question, "Should I adjust my plan?"
  5. Update and monitor the plan regularly and adjust spending or other parts of the plan as needed to keep risk in a reasonable range.”

https://www.kitces.com/blog/retirement-planning-right-capital-market-assumptions-cma-spending-limit-variables-advisor-behavior-client-needs/

We agree with Mr. Fitzpatrick that good financial planning in retirement typically doesn’t involve a one-time only set-it-and-forget it decision, and it is important to establish an ongoing planning process to keep household spending on track. We note that Mr. Fitpatrick’s suggested ongoing planning process steps outlined above are essentially the same as the seven-step process we advocate in our website

Simple “3M” Process for planning in retirement.

The gist of our recommended process can more simply be described using the following 3Ms:

  1. Measure your funded status (ratio of the present value of household assets to the present value of household spending liabilities) using reasonable static assumptions about the future and granular expectations of future spending (recurring and non-recurring)
  2. Monitor your funded status from year to year to observe the trend, and
  3. Make adjustments in your household assets or spending liabilities whenever your Funded Status falls outside a reasonable range (guardrails)

Why the Actuarial Approach is superior to Monte Carlo models typically used by financial advisors for ongoing planning in retirement

Most Monte Carlo models in use today determine the probability that a household can spend $X per year (real dollars) for the expected lifetime of the household based on stochastic assumptions about the future (including assumptions about returns and standard deviations associated with certain asset classes). Unfortunately, most households don’t spend X real dollars every year. Most households have many non-recurring expenses that they will incur over their retirement period. Unlike typical Monte Carlo models, the Actuarial Financial Planner (AFP) can accommodate most of the granular aspects of retirement spending including:

  • non-recurring expenses like cars, home-remodeling projects, long-term care expenses, temporary support payments for children, vacations for a limited period of years, etc.
  • classification of essential expenses vs. discretionary expenses
  • different rates of future increases for different types of expenses, and
  • reduction in expenses upon the first death within the household

It is this granularity that makes the AFP a superior tool for use in the dynamic process anticipated by Mr. Fitzpatrick and the process we advocate. The ability of Monte Carlo modeling to provide probabilities of success through the use of stochastic assumptions has more value for clients and households who prefer the static set-it-and-forget it approach or who are looking to select a desired asset mix. It does not provide as much value when used as part of a dynamic process. This is why most actuaries (like Social Security actuaries and pension plan actuaries) use models employing static assumptions like the AFP model in their dynamic processes. 

Conclusion

If you think it is important to make adjustments in retirement to avoid underspending or overspending, then we advocate the “3M” process outlined above using the Actuarial Financial Planner (AFP) model available for free in our website.